Multi-phase segmentation of the coronary artery in X-ray angiograms

Abstract

An algorithm for segmenting coronary artery from X-ray angiograms as the computer-aided analysis is presented in this paper. It is a fundamental issue in medical imaging system. When a surgeon operates the percutaneous coronary intervention, he references a structure of coronary artery of X-ray angiogram. Therefore, it is important to clearly recognize the coronary artery structure for the correct operation. The artery segmentation by the proposed method makes it easy to understand structure of main coronary artery. This paper describes the algorithm of four phases to segment the coronary artery: a) filtering out noises of X-ray angiogram and enhancing the vessel boundaries, b) automatic segmentation of the main coronary artery using Canny edge detection and adaptive region growing, c) interactive segmentation, to correct unresolved regions, and finally d) compensation of the coronary artery using the morphology method. We have applied our approach to X-ray angiograms of normal and occluded arteries and the results have been validated by clinicians.